Comparison of Daas, Iaas, Paas, Saas, Faas, Baas, Caas

Brief

Recap regarding *aas

• Daas

The abbreviation of Desktop as a service or Data as a service.

• Desktop as a service

The service provides PC's desktop functionalities through internet.

• Data as a service

It provides data management system via internet.

• Iaas

The abbreviation of Infrastructure as a service.

The service provides physical infra like server body, unit, room or etc. Users have to configure layer above OS.

e.g.

AWS EC2

GCP CE

MS Azure

• Paas

The abbreviation of Pratform as a service.

The service provides what is same as Iaas, and also provides OS. Users have to configure layer above FW.  *FW in paas is often limited by vendors.

e.g.

Google App Engine

Herok

Cloud Foundry

• Saas

The abbreviation of Software as a service. Basically accessed by users a thin client via a web browser.

e.g.

Gmail

SAP

Saleceforce

• Faas

The abbreviation of Function as a service.

The service provides small functionalities via webAPI. Mainly for application developers.

e.g.

AWS Lamda

Google Cloud Function

• Baas/MBaas

The abbreviation of Backend as a service/Mobile backend as a service. It provides

functionalities for mobile application development.

e.g.
Fii
Firebase

• Caas

The abbreviation of container as a service

e.g.

GKE

ECS

PKS

Proxy and Reverse Proxy

• Proxy

- Server installed between intranet and internet.

- Client PCs access to internet through proxy.

- Web servers could see only proxy IP address as request originated.

- It is called as forward proxy as well.

• Reverse Proxy

- Server installed in front of multiple web servers.

- Receives http requests instead of web servers.

- It has L7 load balancer functionality and distributes http requests.

- It has cache server functionality as well. Cache web contents like html, css, javascript, xml or etc and return them instead of web servers only if they were not changed after cached.

Performance Test Terms

• throughput

the amount of material or items passing through a system. In general, throughput is the below.

• tps transaction per sec. definition of transaction is different with the projects.
• PV/s  page view per sec

• elapsed time

time that elapses before all the response returns after request

• latency

time that elapses before the first response returns after request

How to make SPA stronger against SEO

Brief

Considering regarding how to make SPA stronger against SEO

Issues

If introduced SAP, the below issues could be considered.

• It's not obvious if web contents updated dynamically is evaluated or not in accurate.
• Google web crawler does not evaluate JS

How to solve

• SSR(Server Side Rendering):  Using node.js or etc.., and execute front end process in server side.
• Prerendering: Using cui browser, and return contents rendered in server side in advance.

kubernetes begining

Brief

By spread of container technology like Docker, it was easy to solve dependencies  of libraries, applications and files among servers. Also, container technology makes it easy to migrate resources.
Next problem is management of containers. If there are just only several containers, managing is not so difficult. But increasing the number of containers, it makes managing more complex.

What's Kubernetes?

Kubernetes is one of the operation tools of container clustering and called as container orchestration tool. It is developed to solve the above container management problems. Google is developing.
It has functions of container operation like auto deploy, storage mount, health check and getting logs or etc...

Numa Architecture and innodb_buffer_pool_size

Event
e.g.
The machine spec
the number of processors: 2
RAM: 32GB

MySQL configuration
innodb_buffer_pool_size = 24GB

If these configurations, Numa architecture divide RAM each16GB. But innodb_buffer_pool_size is over 16GB, so there is possibility that swap occurs.
This event is called as "Swap Insanity".

How to solve

• Deactivate Numa
• Decrease innodb_buffer_pool_size

etc...

Reference
http://d.hatena.ne.jp/fat47/20121121/1353495937

Numa Architecture

What's Numa Architecture
Numa is the abbreviation of "non-uniform memory access".
The objective of Numa is parallelization of *bus.
Numa is an alternative approach that links several small *nodes.

*bus to connect between cpu and RAM.
*node means a pair of processor and RAM.

Features
If Numa architecture is adopted, RAM is divided by the number of bus and accessed.

e.g.
If there are 2 buses on the machine which has 32GB RAM, RAM is divided per 16GB and accessed.